J.Med. Chem. 1981,24,1277-1284 in EtOH (10mL). This was subjected to hydrogenation at 40 psi for 3 h. The suspension was fiitered through Celite and washed with EXOH (50 mL). Evaporation in vacuo gave an oil, which was dissolved in hot EtOH (10mL). EhO was added to the cloud point and the mixture cooled at 5 "C overnight. Filtration gave 5f (717 mg, 75% yield), mp 124-125 OC dec. Anal. (ClZHl8BrN02.HCl) C, H, N. 3-(2'-Bromoethyl)-4-methoxyphenethylamineHydrochloride (5g). This was prepared by the m e procedures as for 12, from 15 oxalate (299 mg, 1 mmol) and PBrs (0.22mL, 2.3 "01). The crude free base was dissolved in EtOH (15mL), and 12 M HCl(O.10 mL) was added. Evaporation gave an oil, which was coevaporated with EtOH (2 x 10 mL) to give a white solid. This was dissolved in hot EtOH (10mL), and EgO (5mL) was added slowly. After cooling to 5 "C overnight, the crystals were filtered and washed with EgO to give 5g (133 mg, 40%), mp 131-132 "C. Anal. (ClZH1&NO.HCl) H, N; C: calcd, 46.70; found, 47.17. Phenethyl Mercaptan Derivatives (6). Method A. To a in EtOH (65mL) was added solution of thiourea (3.80g, 50 "01) 5c (9.25g, 50 mmol) and the mixture refluxed overnight. The solvent was removed in vacuo to give an oil. Trituration with E g o afforded a white solid, which was filtered and washed with EgO to give 6c (11.11 g, 85% yield). Crystallization from hot EtOH, followed by the addition of E g o to the cloud point, gave 6c (5.43 g, 40%), mp 96-98 "C (Table 11). Method B. To EtOH (20mL) and HzO (10mL) was added 5e (1.24g, 3 mmol) and thiourea (456mg, 6.0 mmol). After the mixture was refluxed overnight, 2 N NaOH (6 mL) was added and refluxing continued for 15 min. The mixture was poured into HzO (20 mL) and extracted with CHC13 (3 X 30 mL), dried (MgSOd, fiitered, and evaporated in vacuo to give a clear oil. This oil was dissolved in EtOH (30mL) and 12 N HCl(O.30 mL) was added. Evaporation gave an oily solid, which was triturated in EhO to afford 6e (900 mg, 90%) as a white solid. Crystallization from hot EtOH-EhO gave an analytical sample of 6e,mp 148-149 oc (Table 11). Homocysteine8 (19). Method C. To 30 mL of MeOH in was previously dissolved was added which Na (266mg,11.6 "01) 17 (918mg, 6.0"01). After the mixture was stirred for 15 min, 5d (1.075g, 5.0 mmol) was added, and the reaction was stirred for 1.5 h. Evaporation gave an oily white solid, which was triturated with &O (100mL) and filtered. Evaporation of the filtrate
1277
gave the methyl ester as an oil (900 mg). This oil was dissolved in MeOH (10 mL), 1 N NaOH (10 mL) was added, and the mixture was stirred for 1 h. Evaporation to 10 mL, followed by neutralization to pH 7 with HCl, gave a white precipitate. Filtration afforded 19d (670mg, 50%), mp 192-195 OC dec (Table 111). 5'-Thioadenosines(24). Method D. To a solution of 400 mg (10mmol) of NaOH in 15 mL of HtO was added 1.45 g (5"01) of the isothiouronium salt 6d, and the resulting mixture was heated at 80 "C for 1 h under Nz, at which time 570 mg (2.0 mmol) of 23 was added, and heating continued under N2 for an additional 1 h. The reaction mixture was then cooled, the solution was adjusted to pH 6 with glacial HOAc, and the aqueous supernatant was decanted. The residue was triturated with EhO to give a white solid, which was then crystaUizedfrom E t O H - W , followed by recrystallization from EtOH-Hz0, to yield 610 mg (77%) of 24d. An analytical sample was obtained after two recrystallizetions from CH30H, mp 104-105 OC (Table IV). Method E. To a solution (5 mL) of 2 N NaOH previously purged with Nz was added 285 mg (1 mmol) of 23 and 368 mg (1 mmol) of 6e. After heating under Nz at 70 OC for 4 h, the reaction mixture was cooled and extraded with EtOAc (6X 5 mL), and the dried organic extract was concentrated in vacuo. The resulting residue was dissolved in MeOH, and the desired product, 24e, slowly precipitated from solution: yield 250 mg (43%) of a white solid. Recrystallization from CH30H gave an analytical sample, mp 145-147 OC (Table IV). Sulfonium Salts (3 and 4). Method F. The appropriate thioether (1 mmol) was dissolved in formic acid (2.5 mL) and until the reaction stirred in the dark with Me1 (0.30mL, 5 "01) was judged to be complete by 'H NMR (3-72h). The mixture was poured in ice (10g) and extracted with EhO (3 X 10 mL). The aqueous layer was lyophilized to give the iodide salt of the sulfonium compound, which was dissolved in HzO and passed through an anion resin column in CIOl form. The aqueous eluent was lyophilized to give the sulfonium perchlorate as a white powder (Tables I11 and IV).
Acknowledgment. The authors acknowledge the contribution of Roy Mariuzza in the synthesis and purification of 3c, 4c, and 4d. This research was supported by grants from the US. Public Health Service (MH-18038 and CA10748/ 16359).
Synthesis and Evaluation of Some Stable Multisubstrate Adducts as Specific Inhibitors of Spermidine Synthase Kuo-Chang Tang, Roy Mariuzza, and James K. Coward* Department of Pharmacology, Yale University School of Medicine, New Haven, Connecticut 06510, and Department of Chemistry, Rensselaer Polytechnic Institute, Troy, New York 12181. Received December 1, 1980
A new series of aminopropyltransferaseinhibitors has been designed in which the nucleophilic aminopropyl acceptor is attached to the aminopropyl donor, S-adenosyl-l-(methylthio)-3-propylamine (decarboxylated S-adenosylmethionine), to form a "multisubstrate adduct". In the present case, S - a d e n ~ y l - l , & d i a m i n o 3 - t h i ~(2b) e and the corresponding methylsulfonium salt (3b)have been synthesized. Several compounds of this type were assayed as inhibitors of spermidine synthase, and both 2b and 3b were found to be potent inhibitors of the enzyme. The thioether 2b is the most potent inhibitor of spermidine synthase described to date and is almost totally devoid of inhibitory activity against the closely related aminopropyltransferase, spermine synthase. This type of compound should have use as a specific inhibitor of spermidine biosynthesis in vivo.
The polyamines spermidine and spermine are synthesized by a pair of aminopropyltransferases (APT), spermidine synthase and spermine synthase.' In these reactions, nucleophilic attack by either putrescine or spermidine at a n electrophilic methylene carbon of decarboxyl-
Address correspondence to Rensselaer Polytechnic Institute.
ated S-adenosylmethionine (dcSAM) leads to the formation of t h e polyamine products spermidine and spermine, respectively. Our studies on the mechanism of enzymecatalyzed alkyl-transfer reactions have indicated that the S-adenosylmethionine (SAM) dependent methylase, cat(1) H.G. Williams-Ashman and A. E. Pegg, in "Polyaminee in Biology and Medicine", D. R. Morris and L. Marton, Eds., Marcel Dekker, New York, in press.
0022-2623/81/1824-1277$01.25/00 1981 American Chemical Society
1278 Journal of Medicinal Chemistry, 1981, Vol. 24, No. 11
echo1 0-methyltransferase (COMT), proceeds by a random, sequential kinetic mechanism2 via direct nucleophilic attack of the catechol hydroxyl group on the methyl carbon of SAM: probably involving general-base-catalyzed proton abstraction.* Our earlier studies demonstrated a nonspecific inhibition of SAM-dependent methylases and aminopropyltransferases by the nucleoside products Sadenosylhomocysteine (SAH) and 5’-(methy1thio)adenosine (MTA), respectively.6i6 Therefore, we have initiated a research program aimed at improving the speciticity of synthetic inhibitors of alkyl transfer reactions by incorporating in the inhibitor molecules structural features of proposed enzyme-bound transition states. Reviews by Wolfenden‘ and Jenckss give detailed discussions of the basis for the hypothesis that “transition-state analogues” should be extremely potent inhibitors of enzyme-catalyzed reactions, an hypothesis which is now supported by a considerable body of experimental eviden~e.~ The direct displacement mechanism indicated for the COMT reaction3 suggested to us that the APT reactions might proceed via a similar direct displacement, as shown in 1. This mechanism indicates that nucleophilic attack
Tang, Mariuzza, Coward Scheme I
2
Sa NHR’
I
3
K Sb, R‘ = H c, R’ = CHO
Ox0
4 1
Scheme I1
a,R=H b, R = (CH,),NH,’
on dcSAM by putrescine, catalyzed by spermidine synthase, involves transition state la, whereas nucleophilic attack on dcSAM by spermidine, catalyzed by spermine synthase, involves transition state lb. In this paper we describe the syntheses of 2b and 3b, potential transition-
X
Lx PY
6
x 8
l“Sk
X
LX 9
a,X=H b, X = C1 C,
7
Y
jps9
XA
2
x
X =@oN
d, X = NHCOO-f-Bu e, X = N(C0,-r-Bu), f, X = N, g, X = NH,
3
a, R = C H , b, R = CH(CH,CH,)(CH,),CH,
c, R = CH[(CH,),NH,’](CH,),NH,’ d, R = CH[(CH,),N,I(CH,),N,
state analogue inhibitors of spermidine synthase. The J. K. Coward, E. P. Slisz, and F. Y.-H. Wu, Biochemistry, 12,
2291 (1973). R.W. Woodard, M.-D. Tsai, H. G. Floss, P. A. Crooks, and J. K.Coward, J. Bid. Chem., 255, 9124 (1980). J. 0.Knipe and J. K. Coward, J . Am. Chem. SOC.,101,4339 (1979). J. K. Coward, D. L. Bussolotti, and C.-D. Chang, J. Med. Chem., 17, 1286 (1974). H.Hibasami, R.T. Borchardt, S. Y. Chen, J. K. Coward, and A. E. Pegg,Biochem. J., 187,419 (1980). R. Wolfenden, Annu. Rev. Biophys. Bioeng., 5 , 271 (1976). W. P.Jencks, Adv. Enzymol., 43, 219 (1975).
0
syntheses of the deamino analogues 2a and 3a, predicted to be very poor inhibitors of this enzyme based on substrate structure-activity data? are also described. Preliminary kinetic studies on purified spermidine synthase from rat prostrate reveal that 2b and 3b are potent and specific inhibitors of this enzyme.lOJ1 (9) K.Samejimaand Y. Nakazawa, Arch. Biochem.Biophys., 201, 241 (1980). (10) K.-C. Tang, A. E. Pegg, and J. K.Coward, Biochem. Biophys. Res. Commun., 96,1371 (1980). (11) After this work had been completed, a paper appearedla in which steady-state kinetic studies suggested a double-displacement (“ping-pong”)mechanism for Spermidine synthase isolated from E. coli. However, product inhibition studies were not done nor have these conclusions been confirmed by stereochemical investigations.
Inhibitors of Spermidine Synthase
Journal of Medicinal Chemistry, 1981, Vol. 24, No. 11 1279
Chemistry. The general synthetic approaches to molecules of type 2 and 3 are shown in Scheme I. These routes are well-documented in the literature, including previous work from this lab~ratory.'~J~ Therefore, we investigated synthetic routes to the appropriate RX and RSH for ultimate coupling to 5'-deoxy-5'-thio- and 5'deoxy-5'-chloroadenosine derivatives, respectively. A generalized scheme for the syntheses of the desired alkyl halides and alkyl thioacetates is shown in Scheme 11. For the syntheses of 2b and 3b precursors, conversion of 3octanone to the corresponding secondary alcohol 6a,followed by reaction of the alcohol with PBr,, led to 3bromooctane (sa). The bromide 8a was then converted directly to the desired thiol precursor 9a via bromide displacement with KSAc in dimethyl sulfoxide. Coupling of 9a to 5J-deoxy-5'-chloroadenosine(5a)led to the desired thioether 2b, which could be methylated to 3b with CHJ in HCOOH. Similarly, 8a could be coupled in low yield to the 5J-deoxy-5'-thioadenosinederivative, generated in situ from 4, followed by methylation with CH31/HCOOH to give 3b. Our initial attempts on the syntheses of the diamino analogues 2c and 3c involved the preparation of the key intermediate 1,8-diphthalimido-3-bromooctane(8c)and the corresponding thioacetate 9c. E-Caprolactone was converted to Gchlorohexanoyl chloride via ZnCl2-cataly2ed lactone ring opening in the presence of SOClp Ethylene addition to the acid chloride, catalyzed by AlC13,led to the unstable P-chloro ketone, 1,8-dichloro-3-octanone, which could be reduced with NaBH, to give 1,8-dichloro-3-0ctano1 (6b). Displacement of the chloride atoms of 6b by potassium phthalimide in DMF gave 6c, which could be converted to the desired 3-bromo derivative 8c via the corresponding tosylate 7c. Conversion of the bromide 8c to the thioacetate 9c was again effected by reaction of 8c with KSAc in Me2S0. Unfortunately, neither 8c nor 9c could be successfully coupled to the appropriate adenosine derivatives (Scheme I). The in situ generation of 5'deoxy-5'-thioadenosine from 4, followed by addition of the secondary bromides 8, led to large amounts of the adenosine disulfide,13J4which, in the case of the least complex bromide, 8a,led to low yields of the desired 5'-thioether, even when the reaction was run in the absence of oxygen. In the case of the attempted coupling of 8c,an additional problem was encountered, namely, apparent attack of the thiolate anion on the phthalimide carbonyl groups. Similarly, attempted coupling of the thiolate generated in situ from 9c led to partial cleavage of the phthalimide group under the basic reaction conditions. The use of the tert-butoxycarbonyl (Boc) group to protect the amine functions offered the advantage of base stability, which presumably would avoid the problems encountered with the phthalimide derivatives just described. The introduction of the Boc function is generally accomplished by derivatization of the free amine, which was not appropriate for our work. However, the recent description of the imidies HN(Boc)(C02CH3)and HN( B o c ) led ~ ~us ~ to investigate the possible use of the Boc group in these syntheses. Conversion of 1,8-dichloro-3octanol to either the 1,8-[N(Boc)(C02CH3)12or 1,8-[N-
( B O C ) derivatives ~]~ was effected by reaction of the intermediate 1,8-diiodo-3-octanol with an alkali salt of the appropriate imide in DMF. Base hydrolysis of the methyl ester function in the l,8-[N(Boc)(C02CH3)]2derivative led to the unstable carbamic acid, which spontaneously decarboxylated to give 6d. This material could be converted to l,&diamino-3-octanol(6g) by treatment with CF3COOH. However, all attempts to obtain either the tosylate 7d or bromide 8d failed, presumably due to facile intramolecular cyclization (eq 11, as previously demonstrated
(12) V. Zappia, G.Cacciapuoti, G. Pontoni, and A. Oliva, J. Biol. Chem., 255, 7276 (1980). (13) J. K.Coward, N. C. Motola, and J. D. Moyer, J. Med. Chem., 20,500 (1977). (14) G. L. Anderson, D. L. Bussolotti, and J. K. Coward, J. Med. Chem., preceding paper this issue. (15) C. T.Clarke, J. D. Elliott, and J. H. Jones, J. Chem. SOC., Perkin Trans. 1, 1088 (1978).
with other carbamates under neutral conditions.lB The loss of -50% of the tert-butyl in the recovered product is consistent with the reaction shown in eq 1. The N,N,N',N'-tetrakis(tert-butoxycarbonyl)-3-hydroxy-1,8-octanediamine (6e) lacking the free NH of the Boc derivative 6d should lead to a more stable tosylate (7e)or bromide (7f). Although it proved impossible to obtain 6e as a distillable or crystalline product, a chromatographically and spectrally pure oil was obtained and was converted to the bromide 8e via the tosylate (7e). Both of these materials were viscous oils which were pure by spectral and chromatographic criteria but which defied all crystallization attempts. Coupling of 8e to the EiJ-thioadenosine derivative, generated in situ from 4, led to the totally blocked adenosine thioether precursor of 3c, isolated as an amorphous solid. Methylation of this material with CHJ in HCOOH led to the totally deblocked sulfonium salt 3c. Although this route to 3c was effective in obtaining the target compound, the inability to obtain distillable or crystalline synthetic intermediates in the conversion of 6b to 8e led us to seek an alternative synthetic strategy. The azide function appeared to be a plausible amine precursor for this work, and conversion of l,&dichloro-3-octano1(6b) to the corresponding 18-diazido derivative 6f was accomplished using NaN3 in DMF, containing catalytic amounts of LiI. Tosylation of the alcohol gave 7f, which could be converted to the 1,8-diazido-3-(acetylthio)octane (9f). Coupling of the thiol, generated in situ by base hydrolysis of 9f, to 5J-deoxy-5J-chloroadenosine (5a) gave the diazidooctylthioadenosine precursor 2d. Reduction of the azide functions with triphenylphosphine/pyridinegave the desired diamino thioether 2c. Surprisingly, attempted methylation of 2c to give 3c failed to yield appreciable amounts of the diaminosulfonium salt. Unreacted 2c was the major recovered product, and it seemed possible that in HCOOH the positively charged amino functions of 2c might lead to electrostatic repulsion(s) in the transition gave state leading to 3c. Treatment of 2c with (BOC)~O the 1,8-(B0c)~derivative which, on treatment with (16) J. K. Coward and R. Lok, J. Org. Chem., 38, 2546 (1973). (17) (a) J.Baddiley and G. A. Jamieson, J. Chem. SOC.,1085 (1955). (b) T.Nielson, T.-Y. Shen, and W. V. Rayla, French Patent 589-694;Chem. Abstr., 74,~126013~ (1971). (18) J. Cason and J. S. Correia, J. Org. Chem., 26, 3645 (1961).
1280 Journal of Medicinal Chemistry, 1981, Vol. 24, No. 11 Table I. Inhibition of Spermidine Synthase and Spermine Synthase by 2 and 3 concn, spermidine spermine synthase no. PM synthase 89 2b 25 102 100 104 78 2C 10 3 25 3 90 50 1 100 1 85 3b 10 98 94 25 99 88 50 93 65 100 80 44 250 56 25 3c 5 84 10 68 93 25 31 50 16 60 100 8 41 250 4 14 a Enzyme inhibition data presented as percent of drugfree control.
CH,I/HCOOH, rapidly lost both Boc groups but resisted methylation of the sulfur atom. The only explanation remaining would seem to involve the less rapid hydrolysis of the N(Boc), group vs. the NHBoc group, thus allowing methylation of the neutral thioether derived from 9e to occur prior to hydrolysis of the Boc groups. Thus, synthesis of 3c would appear to require use of the N(Bo42 protecting group in the intermediates. Biochemical Results As noted several years ago in this laboratory, severe substrate inhibition by dcSAM is observed with spermidine synthase.13 This makes the kinetic analysis of inhibition studies with thisenzyme somewhat more complicated than in a normal system. We are currently carrying out more extensive kinetic studies on spermidine synthase. However, preliminary kinetic data (Table I)l0reveal that 3c is a reasonably potent inhibitor (Im = 15 HM)of spermidine synthase and a much less effective inhibitor (1%= 80 pM) of spermine synthase. The deaminosulfonium salt 3b is a poor inhibitor (Im = 250 pM) of spermidine synthase but comparable to 3c as an inhibitor of spermine synthase and similar to other sulfoniums, such as 3a and SAM,against the latter enzyme. Most encouraging, however, for the general approach being investigated are the data with 2c, which is an extremely potent inhibitor (1%= 0.4 ph4) of spermidine synthase, while exhibiting almost no inhibitory activity against spermine synthase. Further kinetic studies are being pursued to investigate the nature of the enzyme-inhibitor interaction more completely. However, these preliminary data suggest that transition-state analogue inhibitors such as 2c should be very useful for studying enzyme-catalyzedalkyl transfer reactions, notably aminopropyltransferases.
Experimental Section All chemicals were of reagent quality and used without further purification with the following exceptions: pyridine and N f i dimethylformamide (DMF)were dried over potassium hydroxide pellets and distilled, 2-butanone was distilled over CaS04prior to use, dimethyl sulfoxide and hexamethylphosphoric triamide (HMPA) were distilled over calcium hydride prior to use, commercial tosyl chloride was recrystallized from petroleum etherbenzene, thionyl chloride was freshly distilled prior to use, and Potassium thicmethanol WBB kept over molecular sieves acetate was triturated with dry 2-butanone several times and dried in vacuo. Melting points were taken on a Mel-Temp capillary melting point apparatus and are uncorrected. Nuclear magnetic
(a).
Tang, Mariuzza, Coward resonance (NMR) spectra were recorded on a Varian T-60 spectrophotometer, IR spectra were measured using a PerkinElmer 237-B spectrophotometer,and UV spectra were measured using a Cary-15 or Perkin-Elmer 552 spectrophotometer. Elemental analyses were performed by Baron ConsultingCo.,Orange, CT. Where analyses are indicated by symbols of the elements, the analytical results were within f0.4% of theoretical values. Thin-layer chromatography (TLC) was performed using either silica gel F-254 (EM Reagents) platea or Eastman cellulw platea with floreacent indicator. Silica gel plates with florescentindicator for preparative chromatography were from Analtech (silica gel GF). Solvent systems used for TLC were as follows: (a) 1BuOH/HOAc/H20, 12:3:5 (BAW); (b) 5% aqueous Na2HP04; and (c) CHC13/MeOH, 4:l. High-performance liquid chromatography (HPLC) was performed using a Whatman ODS-2 reverse-phase column. 5’-Deoxy-5’-chloroadenosine(Sa) was prepared by a modification of the literature procedure.l3J4 All compounds had spectral properties (NMR, JR,and W) consistent with their assigned structures. All adenceyhulfonium compounds (3) (A, 259 nm) decomposed rapidly in 0.1 N NaOH to give 268 nm).Ig adenine (A, 5’-Deoxy-5’-chloro-2’,3’-isopropylideneadenosine (5b). Thionyl chloride (19.3 g, 19.5 mL, 162 “01) was added dropwise to 150 mL of ice-cold dry hexamethylphosphoric triamide ( W A ) under a nitrogen atmosphere. To the resulting cooled solution of 2’,3’-isopropylideneadenosinein was added 18.6 g (60 “01) portions. When the adenosine was added the color of the solution changed im~~ediately from pale yellow to orange and then reddish. After 5 h of stirring at ambient temperature, the reaction mixture was poured into a well-stirred mixture of icewater (900 g), and the reaction flask was rinsed with water. The aqueous solution thus obtained was adjusted to pH 9 with concentrated NHIOH to give a white precipitate. The precipitate was collected and redissolved in chloroform. The chloroform solution was decolorized with charcoal to give a pale yellow solution, which was then poured into petroleum ether to give 17.2 g (88%)of a white precipitate with mp 278 “C dec. This material was shown to be homogeneous on TLC (siliq EtOAc/CHCqt %l)and HPLC (50% aqueous methanol, ODs-1): NMR (CDCla) 6 1.4 (3 H, s), 1.6 (3 H, s), 3.7 (2 H, dd, Ha,, J = 6 and 5 Hz), 4.46 (1H, dt, H49 J = 2 Hz), 5.11 (1H, dd, Hy, J = 2 and 7 Hz), 5.48 (1H, dd, Hy, J = 2 and 7 Hz), 6.1 (1 H, d, HI,, J = 2 Hz), 6.6 (2 H, br s, NH2), 7.86 (1H, s, H.j, 8.25 (1H, s, H8). Anal. (C13H16C1Ns03)C, H, N. 5‘-Deoxy-S‘-chlo~N’-formyI-~~-~p~py~~~ (5c). 5b (222 g, 6.8 “01) was dissolved in 20 mL of acetieformic anhydride, and 0.8 g (7.55 “01) of anhydrous sodium carbonate was added. The resulting solution was allowed to stir at ambient temperature for 8 h, and the reaction was monitored by TLC (silica gel; EtOAc/CHC13,91). At the beginning, CO2 evolution occurred and a slightly cloudy solution was obtained, but this became clear at the end of the reaction. The mixture was then concentrated to almost dryness under reduced pressure, and the residue was dissolved in 50 mL of CHC13,which was then washed with HzO (2 x 40 mL), saturated aqueous NaHC03 (2 x 40 mL), and H20 (40 mL) and dried over MgSO,. After the solvent was removed 2.1 g (85%) of pure 5c was obtained mp 230 “C dec; NMR (Me2SO-de)6 1.28 (3 H, s), 1.49 (3 H, s), 3.82 (2 H, d, H ~ TJ ,= 6 Hz), 4 41 (1 H, dt, Hdt, J = 2 and 6 Hz), 5.11 (1H, dd, Ha., J = 2 and 7 Hz), 5.54 (1 H, dd, Hr, J = 7 Hz), 6.41 (1H, d, Hi!, J = 2 Hz), 8.64 (1 H, S, Hz), 8.67 (1 H, 8, Hd, 9.87 [ l H, d, -NHC(=O)-, J = 9 Hz], 11.24 (1H, d, C H O , J 9 Hz). Anal. ( C ~ ~ H I & ~ N SC,OH, ~ )N. 5’-Deoxy-5‘-(thioacetyl)-Ns-formyl-2’,3’-isopropylideneadenosine (4). Sc (0.5 g, 1.42 mmol), previously triturated potassium thioacetate (0.485g, 4.26 mmol), and a catalytic amount of anhydrous LiI were dissolved in 35 mL of 2-butanone, and the resulting solution was refluxed for 5 h. After cooling, the dark brown solution was filtered with the aid of Celite, and the fiitrate was concentrated under reduced pressure with the bath temperature below 30 OC. The residue thus obtained was taken into 40 mL of chloroform and washed with H20 (4 X 40 mL) and dried (19) R. T. Borchardt, J . Am. Chem. SOC.,101,458 (1979), and references therein.
Journal of Medicinal Chemistry, 1981, Vol. 24, No. 11 1281
Inhibitors of Spermidine Synthase
over MgSO,. After the solvent was removed, 280 mg of crude product was obtained, which, after recrystallization from CHC13-Etz0, gave 235 mg (42.1%) of pure 4, which exhibited spectral properties and melting point identical with that of an authentic compound prepared by the method published previOUS~Y.~~
3-0ctanol(6a). To a well-stirred solution of 19.2 g (0.15 mmol) of 3-octanone in 270 mL of 95% ethanol, cooled in an ice bath, was added in portions a solution of 3.9 g (0.103 m o l ) of sodium borohydride in 27 mL of water. Ammonium hydroxide (15 M, 27 mL) was added, the ice bath was removed, and stirring was continued at room temperature for 3 h. The reaction mixture was concentrated to near dryness, and the residue was partitioned between 250 mL each of CHC1, and HzO. The organic layer was separated and the aqueous layer was extracted with CHCl3(2 X 200 mL). The combined extracts were washed with 5% HCl(350 mL) and saturated NaCl solution (350 mL) and dried over MgSOI. After the solvent was removed, the liquid residue was distilled under reduced pressure [bp 86-87 "C (24 torr), lit.l8 bp 69.5-70.4 "C (7 torr)] to give 15.41 g (79.02%) of pure 6a: NMR (CDC13) 6 0.93 (6 H, t, CH3), 1.37 (10 H, br m, CHz), 2.35 (1 H, br s, exchangeable with DzO and shift to low field in pyridine, OH), 3.47 (1 H, m, >CHO-). 3-Bromoctane (8a). To 8 mL (23g, 85.3 mmol) of phoephorus tribromide was added dropwise 15 g (115.4 mmol) of 3-octanol (6a) over a 0.5-h period. The resulting solution was heated at 100 "C (oil bath) for 2 h, after which time the reaction mixture was cooled and poured into 200 mL of ice-water, which was extracted with CHC13 (3 x 120 mL). The combined organic extracts were washed with 5% Na&O3 (2 x 200 mL), H20 (200 mL), saturated aqueous NaHC03 (2 X 200 mL), HzO (200 mL), and saturated aqueous NaCl(200 mL) and dried over MgSO,. After the solvent was removed under reduced pressure, the residue was distilled [bp 78-80 "C (18torr), lit.18 bp 84.4-85.1 "C (20 torr)] to give 17.5 g (78.58%) of pure 8a, which exhibited a C-Br band at 797 cm-' in the IR spectrum:18 NMR (CDC13) 6 1.0 (6 H, t, CH,), 1.31 (6 H, br s, CHJ, 1.75 [4 H, q, CHzC(Br)CHzI,3.90 (1 H, m, >CHBr). S-Adenosyl-3-(methylthio)octane(3b). Route A. From 5'-Deoxy-5'-chloroadenosine(5a). 3-(Thioacetyl)octane (Sa). To 6.2 g (54.3 mmol) of previously triturated potassium thioacetate in 60 mL of dry MezSO was added 7.0 g (36.27 mmol) of 3bromoctane @a),and the resulting solution was stirred overnight at ambient temperature. The reaction mixture was then poured into 500 mL of HzO and extracted with chloroform (2 X 250 mL). The combined CHC13 extracts were washed with H20 (2 X 250 mL) and saturated aqueous NaCl(250 mL) and dried over MgSO,. After the solvent was removed, 6.8 g ( 100%) of crude product residue was distilled [bp 94-95 "C (22 torr)] to give 6.1 g (89.7%) of pure Sa: NMR (CDC13)6 0.67-1.10 (6 H, m, CH3), 1.10-1.97 (10 H, m, CHz), 2.3 [3 H, s, CH,C(=O)], 3.48 (1 H, quintet, >CHS); IR (thin film) 1689 (>C=O) cm-'. Anal. (Cld-IH,OS) C, H, S. S-Adenosyl-3-thiooctane (2b). 3-Thioacetyloctane(Sa; 480 mg, 2.55 mmol) in 10 mL of dry MezSO was degassed with a stream of nitrogen for 1h, after which time 485 mg (1.7 mmol) of 5'-deoxy-5'-chloroadenosine (5a) was added, followed by 2 mL of 4 M NaOH. The resulting solution was stirred at ambient temperature overnight. The reaction mixture was then poured into 175 mL of HzO to give a milky cloudy solution, which was cooled at -20 "C. After the solution was warmed to -25 "C, a solid was collected by filtration to give 600 mg (88.8%) of crude 2b. The crude product was recrystallized from MeOH-HzO to give 550 mg (81.4%) of pure 2b: mp 77-80 "C; NMR (CD30D) 6 0.5-0.95 (6 H, m, CH3), 0.95-1.76 (10 H, m, CH2),2.48 (1 H, quintet, >CHS), 2.76 (2 H, d, HSt,J = 6 Hz), 3.4-4.37 (2 H, m, Hy and H4,) 5.88 (1 H, d, Hi,, J 5 Hz), 8.06 (1 H, 8, Hg), 8.16 (1H, s, Ha,Hz peak obscured by OH signal (6 4.33-5.1); UV A219,260 nm; TLC Rf 0.93 (celluloee;BAW). Anal. (C,$&,O$) C, H, N, S. S-Adenosyl-3-(methylthio)octane(3b). 2b (160 mg, 0.4 "01) was methylated in 1.5 mL of 88% formic acid with 0.5 mL of methyl iodide. The resulting solution was stirred at ambient temperature, protected from the light, for 3 days, after which time the reaction mixture was partitioned between 50 mL each of ether and HzO. The aqueous layer was separated, washed with ether
-
(3 X 50 mL), and lyophilized to give 150 mg (61.9%) of the iodide
salt: NMR (D2O) 6 0.6-1.5 (12H, broad complex, CH3and CHd, 1.5-2.05 [4 H, broad complex, CHZC(S)CHz],3.66 (1 H, m >CHS+CHS), 4.28 (1H, dt, H4,,J = 6 and 2 Hz), 4.95 (1H, dd, Hy, J = 2 and 7 Hz), 5.43 (1H, dd, Hz, J = 2 and 7 Hz), 5.98 (1 H, d, H1,, J = 2 Hz), 6.21 (2 H, br s, NHz), 7.81 (1H, s, Hz),8.2 (1H, s, Ha). This thioether could be converted to 3b by methylation with CH31 in HCOOH, as described above for route A. The product 3b obtained by route B was identical by NMR, mp, and TLC with that obtained by route A. 1,8-Dichloro-3-octanol(6b).A. 6-Chlorohexanoyl Chloride. Zinc chloride (2.26 g, 16.6 mmol) was added to 77.3 g (667 mmol) of e-caprolactone in an ice bath to give a reddish solution to which 94.0 g (790 mmol) of thionyl chloride was added dropwise. After the addition, the color of the reaction mixture was a dark brown, which gradually became lighter in color as the reaction mixture was heated at 50-60 "C overnight. NMR spectra indicated that all the starting lactone had been consumed by this time. After the excess of thionyl chloride was removed under reduced pressure, the crude product was vacuum distilled [bp 70-72 "C (1.5 torr)] to give 57.49 g (51%)of a clear, colorless liquid NMR (CDC13)6 1.63 (6 H, m, CHz),2.88 [2 H, t, CHzC(=O)C1, J = 3.5 Hz), 3.57 (2 H, t, CHzC1, J = 3.0 Hz); IR (thin film) 1798 (>W) cm-'. Anal. CsHloClzO(C, H). B. l,&Dichloro-3-octanone. A well-stirred, iceaoled solution of 30.35 g (179.6 mmol) of 6-chlorohexanoylchloride in 300 mL of dry CCll was degassed with nitrogen for 0.5 h. AlC13 (26.34 g, 179.6 mmol) was added in portions, and then ethylene was bubbled in at a rate so that no excess ethylene escaped from the reaction vessel. Ethylene was allowed to bubble through the reaction mixture at 0 "C for 2 h and at ambient temperature overnight. The reaction mixture was poured into 800 mL of ice-water and extracted with chloroform (2 X 400 mL). The combined chloroform extracts were washed with saturated aqueous NaHC03 (500mL), HzO (500mL), and saturated NaCl(500 mL) and dried over MgSO,. After the solvent was removed, 15.2 g (43%) of crude l,&dichloroodanonewas obtained, which was used without further purification: NMR (CDCld 6 1.44 (6 H, m, CHz), 2.34 [2 H, t, CHzC(=O), J = 3.0 Hz], 2.85 (2 H, t, CH,C(=O), J = 3.0 Hz), 3.52 (2 H, t, CHzC1, J = 3 Hz), 3.74 (2 H, t, CH2C1, J = 3.0 Hz). C. 1,8-Dichloro-3-octanol (6b). To a well-stirred solution of 15.08 g (77 mmol) of 1,8-dichloro-3-octanonein 15 mL of 95%
1282 Journal of Medicinal Chemistry, 1981, Vol. 24, No. 11
ethanol, cooled in an ice bath, was added a solution of 1.97 g (52 "01) of sodium borohydride in 7 mL of H20. Concentrated NKOH (20 mL) was added, and the resulting solution was stirred at ambient temperature for 1h after the ice bath was removed. The reaction mixture was poured into 350 mL of HzO and extracted with CHC1, (2 X 350 mL). The combined extracts were washed with 5% HC1 solution (300 mL) and dried over MgSO,. After the solvent was removed under reduced pressure, the regidue was vacuum distilled [bp 95-97 "C (0.3 torr)] to give 10.49 g (69%) of pure 6b: NMR (CDClJ 6 1.42 (6 H, br m, CHJ,l.85 [4 H, m, CH&(OH)CH,], 2.24 (1H, br 8,OH, exchangable), 3.33-3.9 (5 H, overlapped multiplet, CHzCland >CHO)iIR (thin film)3356 (OH) cm-'. Anal. (C&Il6Cl20)C, H, C1. 1,8-Diphthalimido-3-octanol (6c). A solution of 4.87 g (45 mmol) of 1,8-dichloro-3-octano1(6b)and 9.97 g (54 mmol) of potassium phthalimide in 40 mL of dry DMF was heated at 100 OC for 2 h. As the reaction proceeded, the potassium phthalimide was slowly drawn into the solution to produce a slightly greenish color, and a fine white precipitate formed, presumably KCl. After cooling, the reaction mixture was partitioned between 300 mL each of CHCls and HzO. The organic layer was separated and washed with HzO (300 mL), 1N NaOH (300 mL), and HzO (300 mL) and dried over MgSO,. After the solvent was removed under reduced pressure, an oily residue was obtained, which crystalked upon trituration with ether. The solid was collected by filtration and recrystallized from MeOH-Et20 to give 6.26 g (61%) of pure 6c as long needle-shaped crystals: mp 144.5-145.5 "C; TLC Rf 0.35 (silicagel; MeOH/CHC13, 1:24); NMR (CDCI,) 6 1.0-2.01 (10 H, complex, CHJ, 2.75 (1 H, br s, OH), 3.254.01 (5 H, m, CHJW C, and >CHO), 7.71 (8 H, br s, aromatic). Anal. (C24H24N205) H,N. 1,8-Diphthalimido-3-(tosyloxy)octane(712). To a solution of 3.0 g (7.1 "01) of 1,8-diphthalimido-3-octanol(612)in 15 mL of dry pyridine, cooled in an ice bath, was added in portions 5.63 g (29.5 mmol) of recrystallized tosyl chloride to give a brownish solution. After -10 min at 0 "C, a white precipitate, presumably pyridinium hydrochloride, formed, and the resulting mixture was stirred overnight at 4 "C. The mixture was then poured into 300 mL of ice-water and extracted with CHC13 (2 X 200 mL). The combined organic extracts were washed with cold 1N HCl(3 x 100 mL), and saturated aqueous NaCl(l50 mL) and dried over MgS04. After the solvent was removed under reduced pressure, a yellow oily residue was obtained, which gave a white solid after trituration several times with ether and cooling to -20 "C for several hours. The solid thus isolated [3.56 g (87%);R, 0.60 (silica gel; MeOH/CHC13, 1:24] was used without any further purification: NMR (CDCl,) 6 1.12-2.10 (6 H, m, CHz), 1.97 [4 H, m, CHzC(OH)CHz],2.40 (3 H,s,CH3),3.35-3.86 (4,H,m, CHzNCHO), 7.71 (8 H, br s, aromatic). Anal. (Cz4HZ3BrN2O4) C, H, N. 1,8-Diphthalimido-3-( thioacety1)octane (9c). The bromide 8c (400 mg, 0.828 mmol) was reacted with 144 mg (1.26 mmol) of previously triturated potassium thioacetate in 5 mL of dry MezSO at ambient temperature. After workup as described previously for 9a,the oily residue was triturated with MeOH to give 278 mg (70%)of 9c as a chromatographically pure solid, mp 120.5-122 "C. The analytical sample was recrystallized from hot MeOH: NMR (CDCl,) 6 1.06-2.1 (10 H, complex, CH2),2.2 [3 H, s, CH3C(=O)J, 3.36-3.9 (4 H, m, CHzNCHS), 3.60 (4 H, m, CH,NCHS, CH2N3);IR (thin film) 2105 (N3), 1687 (>C=O) cm-’; TLC R 0.82 (silica gel; MeOHcalcd, 6.71; found, 7.35. CHC13, 1:4). Anal. (Cl&Il&O) C, N S-Adenosyl-1,8-diazido-3-thiooctane (2d). Pure 9f (688mg, 2.55 mmol) was coupled with 5’-deoxy-5’-chloroadenosine(Sa) by the procedure described previously for 2b to give 541 mg (66.7%) of practically pure 2d, which was recrystallized from MeOH-H20 to give pure 2d: mp 44-46 “C; TLC Rf 0.71 (silica gel; MeOH-CHC13,1:4);HPLC t, 21.3 min (ODS-2,65%aqueous MeOH); NMR (CD30D) 6 0.83-1.9 (10 H, complex, CHz), 2.9-3.53 2.26-2.63 (1H, m, >CHS), 2.63-2.9 (2 H, d, Hs, J = 6 Hz), (4 H, m, CHzN3),3.83-4.4 (2 H, complex, H3, and H,,), 5.91 (1 H, d, Hlr,J = 5 Hz), 8.1 (1H, s, Hz),8.16 (1H, s, Ha), Hr peak obscured by OH signal (6 4.36-4.85); IR (Nujol) 2118 (N3)cm-’; W A, 210,259 nm; TLC R, 0.71 (silica gel; CHC13-MeOH, 4 0 , N
d:
J. Med. Chem. 1981, 24, 1284-1287
1284
0.88 (cellulose; BAW). Anal. ( C & I ~ N l l ON ~ )C: calcd, 45.27; found, 44.72; H: calcd, 5.70; found, 6.61; S: calcd, 6.72; found, 7.27. S-Adenosyl-l,8-diamino-3-thiooctane (2c). S-Adenosyll&diazid0-3-thiooctane (2d; 280 mg, 0.49 mmol) and triwere dissolved in 1mL of phenylphosphine (420 mg, 1.6 "01) dry pyridine, and the resulting solution was kept at ambient temperature with stirring for 1h, during which time gas evolution (presumably Na) was observed. Ammonium hydroxide (15 M, 300 NL)was then added, and stirring was continued for another 2 h. The excess ammonium hydroxide and pyridine were removed under high vacuum at mom temperature, and the resulting residue wm dissolved in 70 mL of HzO. The aqueous solution was washed with benzene (3 X 50 mL) and ether (3 X 50 mL) and then lyophilized to give 212 mg (82.5%) of free aminonucleoside as a hygroscopic white solid NMFt (D20) S 0.67-1.96 (10 H, br, CH2), 2.3-3.06 (7 H, complex, CH,N
